TrendsRecent methodological advancesallow us to study G-quadruplex struc-tures at higher resolution andthroughput.

Approaches to use G-quadruplexstructures as molecular tools arehighlighted.

Computational and experimentalmethods for G-quadruplex studiesare reviewed.

The works reviewed herein provide

ReviewG-Quadruplexes: Prediction,Characterization, andBiological ApplicationChun Kit Kwok1,*,@,z and Catherine J. Merrick2,*,z

Guanine (G)-rich sequences in nucleic acids can assemble into G-quadruplexstructures that involve G-quartets linked by loop nucleotides. The structuraland topological diversity of G-quadruplexes have attracted great attention fordecades. Recent methodological advances have advanced the identificationand characterization of G-quadruplexes in vivo as well as in vitro, and at a muchhigher resolution and throughput, which has greatly expanded our currentunderstanding of G-quadruplex structure and function. Accumulating knowl-edge about the structural properties of G-quadruplexes has helped to designand develop a repertoire of molecular and chemical tools for biological appli-cations. This review highlights how these exciting methods and findings haveopened new doors to investigate the potential functions and applications ofG-quadruplexes in basic and applied biosciences.

unique insights to explore the biologi-cal roles and uses of G-quadruplexesin basic and applied research.

1Department of Chemistry, CityUniversity of Hong Kong, KowloonTong, Hong Kong SAR, China2Centre for Applied Entomology andParasitology, Faculty of NaturalSciences, Keele University, Keele,Staffordshire, UK@Twitter: @kitkwok6zLab website: http://www.kitkwok.weebly.com (C.K. Kwok); https://www.keele.ac.uk/lifesci/people/catherinemerrick/ (C.J. Merrick).

*Correspondence:ckkwok42@cityu.edu.hk (C.K. Kwok)andc.merrick@keele.ac.uk (C.J. Merrick).

Importance of G-Quadruplex Structures in BiologyNucleic acid structures are fundamental to cellular function and the regulation of diversebiological events [1]. DNA and RNA sequences can fold into myriad structural motifs, suchas duplexes, hairpins, triplexes, pseudoknots, and G-quadruplexes, to assemble the functionalstructural conformation for their precise biological roles in specific cellular environments [1].Interestingly, guanine (G)-rich sequences can self-associate into stacks of G-quartets(Figure 1A) to form complex structural motifs known as G-quadruplexes [2] (Figure 1B). G-quadruplexes are of growing interest in chemistry and biology, largely due to their peculiar anddiverse molecular structures, which include parallel and antiparallel topologies (see Glos-sary; Figure 1B). Recently, G-quadruplexes have been reported to have critical regulatory rolesin biological processes, including but not limited to DNA replication, transcription, and transla-tion [3,4] (Figure 1C–E), providing new and important mechanisms for controlling gene expres-sion and genome stability. By understanding the principles of how G-quadruplex structuresmediate gene expression in cells, one can harness their chemical and biochemical properties toaid in developing novel biological applications.

In this review, we first illustrate how G-quadruplexes can be used in biological applications.Next, we present the classical methods used to predict and identify G-quadruplex structures.Then, we highlight new molecular and chemical tools that enable detection of G-quadruplexesin cell imaging, followed by innovative next-generation sequencing (NGS) techniques thatmap G-quadruplex structures on a genomic or transcriptomic scale. Novel biological insightsthat have resulted from these studies, and the current limitations of these methods, arediscussed. Lastly, we present our perspectives on future advances and challenges towarda more complete understanding of G-quadruplex structure–function relationships in vivo, which

Trends in Biotechnology, October 2017, Vol. 35, No. 10 http://dx.doi.org/10.1016/j.tibtech.2017.06.012 997© 2017 Elsevier Ltd. All rights reserved.

GlossaryAptamer: a biological molecule –

usually a peptide or oligonucleotide –

that binds to a specific target suchas a protein or small molecule.Oligonucleotide aptamers (which mayform G-quadruplexes) can begenerated by combinatorial nucleicacid library screening, SELEXexperiment, and other methods.Chromatin immunoprecipitation(ChIP): a technique to locateproteins – and also DNA motifs suchas G-quadruplexes – in nativechromatin. Chromatin isformaldehyde fixed, extracted fromcells, fragmented, and treated withan antibody to the entity of interestto isolate associated DNA fragments.These are then identified bysequencing (ChIP-seq) or byhybridization to a microarray (ChIP-on-chip).Intermolecular G-quadruplex:formed from runs of guanines onmore than one DNA strand, or froma hybrid of DNA and RNA strands.See Figure 1B.Intramolecular G-quadruplex:formed from a single DNA strand,which bears four runs of guanineresidues in close proximity. SeeFigure 1B.Light-off probe: undergoesfluorescence quenching upon bindingto its target.Light-up probe: displays enhancedfluorescence upon binding to itstarget.Next-generation sequencing(NGS): modern, high-throughputsequencing techniques, such asIllumina, Ion Torrent, and 454, all ofwhich produce sequence dataconcurrently on a genomic/transcriptomic scale in the form ofmillions of short sequence fragments(usually <1 kb).Parallel and antiparalleltopologies: a parallel G-quadruplexhas all of the guanine-bearingstrands in the same 50/30 polarity,necessitating linking by ‘propeller-type’ loops that run top-to-bottom ofthe folded motif. In an antiparallelquadruplex, the strands do not allhave the same polarity, and thus thelinking loops can be at the top orbottom of the folded motif. SeeFigure 1B.SELEX: ‘systematic evolution ofligands by exponential enrichment’. Atechnique for generating highlytarget-selective oligonucleotides with

will facilitate the potential development of a new set of G-quadruplex-based biologicalapplications.

G-Quadruplexes As Molecular Tools for Biological ApplicationsBeing versatile in nature, G-quadruplexes have been identified on many occasions for thedevelopment of molecular tools binding to diverse classes of targets (Table 1). It is desirable touse G-quadruplex-containing aptamers as therapeutic and diagnostic agents for diseasessuch as cancers [5–7], as they are thermodynamically and chemically stable, resistant to manyserum nucleases, and have a low immunogenicity and good cellular uptake [8].

Among the G-quadruplex-containing aptamers reported to recognize proteins and enzymes(Table 1), one of the well-studied examples is thrombin-binding aptamer (TBA) [9], identified byan in vitro selection process called systematic evolution of ligands by exponentialenrichment (SELEX; Figure 2A), which binds to the exosite I of human thrombin with highaffinity and selectivity. The crystal structure of TBA revealed that the DNA G-quadruplex is inantiparallel topology with two-quartet planes [10], and data suggested that the presence ofthrombin induces the TBA to fold into a G-quadruplex conformation for binding [11]. Recentstudies on TBA and variants suggest that they can be used for biosensing, with nanomolaraffinity to thrombin [12,13].

Besides proteins and enzymes, specific G-quadruplex-containing aptamers have also beenfound to recognize small molecules (Table 1). SELEX was used to identify an aptamer called‘spinach’ that recognizes 3,5-difluoro-4-hydroxybenzylidene imidazolinone, and producesfluorescence upon ligand binding [14]. Recent crystal structural studies showed that the ligandstacked with the unique G-quadruplex [15] (Figure 2B). Notably, the G-quadruplex was foundto be in a special structural scaffold in the spinach aptamer that is essential for the ligandbinding and fluorescence [15]. Since then, the spinach aptamer has been modified and appliedfor live cell imaging to track different biomolecules [16,17]. In addition, the sequence require-ment for the fluorescence of spinach RNA aptamer was extensively studied [18], and severalother fluorescent RNA aptamers, such as ‘mango’ aptamer and other variants of ‘spinach’,have been shown to contain G-quadruplex structures [19–22], providing structural insights forfuture design and development of brighter fluorescent RNA for imaging purposes.

In one particularly versatile application of G-quadruplexes as biosensors, the G-quadruplex-containing aptamer PS2.M, initially selected via SELEX, was shown to possess catalyticproperties upon binding with hemin [23] (Figure 2C). This G-quadruplex–hemin complexcan be used to mimic the activity of horseradish peroxidase [23] by catalytically oxidizingcolorimetric substrates such as 3,30,5,50-tetramethylbenzidine and 2,20-azino-bis(3-ethylben-zothiazoline-6-sulfonic acid) upon hydrogen peroxide addition (Figure 2C). The sequence andstructure requirements for the catalytic properties were tested comprehensively to verify theimportant role of the G-quadruplex in this peroxidase-mimicking system [24–27], which hassince been developed into numerous applications including the detection of cations, organicmolecules, proteins, nucleic acids, and others [28].

Another notable G-quadruplex application is quadruplex priming amplification (QPA). Thestrategy involves the spontaneous dissociation of DNA duplex and formation of DNAG-quadruplex upon primer extension reaction. The G-quadruplex is then detected via fluo-rescence signal from 2-aminopurine (Figure 2D) [29], which substitutes the T at the first loop ofthe dG3T G-quadruplex, giving a fluorescence increase upon G-quadruplex formation. TheQPA can be employed for both linear [29] and exponential [30] signal amplification modes todetect target sequences of interest. For example, QPA can be coupled with linear nickingamplification [31], which allows detection of target sequences as low as the 10 fM [32]. This

998 Trends in Biotechnology, October 2017, Vol. 35, No. 10

strong binding affinity from a libraryof random sequences via repeatedrounds of binding to the targetligand, washing, elution, reversetranscription (for RNA aptamer), andPCR amplification.SHAPE: ‘selective 20-hydroxylacylation analysed by primerextension. SHAPE is used todetermine RNA secondary structuresby treating RNA with an acylationreagent that selectively acylates theflexible (unpaired) nucleotides of theRNA at the 20-hydroxyl (20-OH)group. These modifications can stallreverse transcriptase and thusprovide an electrophoresis-based orNGS-based readout of nucleotidereactivity, which can then be used toinfer RNA structure.

approach paves the way to detect low-abundance nucleic acid molecules, such as pathogenicDNA, for diagnostic applications. In QPA, the readout is fluorescence signal from 2-amino-purine or 3-methylisoxanthopterin; however, it might be interesting to see if it can be performedwithout the need of exogenous fluorophores, as several recent studies have reported theintrinsic fluorescence of the G-quadruplex in DNA and RNA [33–38].

Predicting G-Quadruplexes Using Computational MethodsClassically, a nucleic acid sequence containing four runs of at least three guanines, separated byshort stretches of other bases, can potentially fold into an intramolecular G-quadruplex, so thepotential to form these motifs can be predicted from primary sequence. In an intramolecularG-quadruplex, the guanine runs all occur on the same strand of DNA, whereas in an intermo-lecular G-quadruplex they occur on both the sense and antisense strands (Figure 1B). Forintramolecular quadruplexes, many algorithms have been published over the past decade topredict the potential formation of G-quadruplexes directly from DNA sequence, including Quad-Parser [39], QGRS Mapper [40], G4P Calculator [41], QuadBase [42] and most recently, G4Hunter [43]: their relative features are reviewed in [44]. Some of these algorithms simply seeksequences bearing four tracks of three guanines in close proximity; others take into accountadditional factors that can influence G4 folding (discussed below) such as the length and nature ofloops between the tracks of guanine. Predicting intermolecular G-quadruplexes is somewhatmore complex because it requiresconsiderationofbothDNAstrands [45,46],and algorithms havealso been developed specifically for RNA [47] which, being single-stranded, can adopt myriadcompeting conformations besides G-quadruplexes [48].

Using such algorithms, it is of theoretical interest to predict how many putative G-quadruplexsequences (PQSs) a genome would be expected to contain at random – and therefore whetherthese sequences are over- or under-represented in real genomes. However, this remains anontrivial problem because any simple model is rendered highly inadequate by variable genomecomposition (i.e., the overall percentage of guanines in a genome may not be distributedequally throughout) and by biased base dyad frequencies (i.e., the nonrandom likelihood thatany G will be followed by a G, C, T, or A) [39,49].

The consensus sequence for PQSs has traditionally been G3N1–7 G3 N1–7 G3 N1–7 G3 [39], but itis increasingly recognized that this consensus does not accurately predict all the PQSs in agenome [50]: motifs with larger loops, nonguanine bulges, etc. may also form G-quadruplexes.Loops as large as N = 30 can support G-quadruplex formation in vitro [51] and most predictivealgorithms permit a user-defined loop length. Short loops are, however, a major factor inG-quadruplex stability [52,53] and some algorithms incorporate this, together with otherfactors, into a sliding score for G-quadruplex propensity and stability, rather than a binaryprediction [41,43,54]. The picture is further complicated by recent evidence that ‘bulged’G-quadruplexes can occur in vitro (with a nonguanine base interrupting a 3-guanine tracksequence) [50,55], and also that two instead of three guanine quartets can suffice, particularlyin RNA, giving rise to G2 Nx quadruplexes [56]. Since no predictive algorithm is perfect for allpurposes, users must define their parameters appropriately and balance the chances of falsepositives against false negatives. It is then important to confirm in silico predictions via the invitro and/or in vivo methods discussed below.

Identifying and Characterizing G-Quadruplexes Using Biophysical andBiochemical MethodsA number of experimental methods have been developed to support the computationalprediction of G-quadruplexes. These methods can be broadly defined into two classes:biophysical and biochemical methods (Table 2).

Trends in Biotechnology, October 2017, Vol. 35, No. 10 999

Figure 1. G-Quadruplex Structure and Biology. (A) Chemical structure of a G-quartet. Potassium ion (K+) sits within the G-quartet for stabilization. G-quartetsstack on each other to form G-quadruplex. (B) Representative topologies of G-quadruplex structures. (C–E) Representative G-quadruplex-associated biology:regulation of (C) DNA replication, (D) transcription, and (E) translation.

Thanks to the unusual structure and folding of the G-quadruplex, one can experimentallyidentify G-quadruplex formation and investigate structural properties using biophysical tech-niques (Table 2). For example, the topology of the G-quadruplex structure can be determinedby monitoring the positive or negative circular dichroism (CD) signals at specific wavelengths[57]. In general, G-quadruplexes with parallel topology (Figure 1B) have negative and positiveCD signals at 240 and 262 nm, respectively, whereas antiparallel topology (Figure 1B) placesthese signals at 262 and 295 nm, respectively. To verify G-quadruplex formation, one shouldalso perform the CD experiments under non-G-quadruplex stabilizing (Li+) and G-quadruplexstabilizing conditions (such as K+ or with G-quadruplex ligands), and scan toward the far-UVregion (�180–230 nm). Likewise, the thermostability of the G-quadruplex structure can beidentified by observing the UV signal at 295 nm [58]. Upon G-quadruplex melting, the UVabsorbance at 295 nm decreases, leading to a hypochromic shift that is a distinctive feature ofG-quadruplex structure. In addition, G-imino protons involved in G-quartets will exhibit adistinct proton chemical shift value of 10.5–12 ppm in NMR. Moreover, certain dyes suchas the benzothiazole Thioflavin T (ThT) and N-methyl mesoporphyrin IX have been found tofluoresce upon binding to G-quadruplexes [59,60], providing light-up structural probes todetect G-quadruplexes in vitro: such chemical tools were recently reviewed [61]. These andother biophysical techniques (Table 2) are widely used under different in vitro conditions to verifyG-quadruplex formation; however, these methods are limited to studying short oligonucleo-tides and thus do not account for the effect of flanking sequences.

1000 Trends in Biotechnology, October 2017, Vol. 35, No. 10

Table 1. Representative List of G-Quadruplex-Containing Aptamers

G4/rG4-containing aptamer Targetsa Aptamer sequences Refs

T40214 STAT3 d(GGGCGGGCGGGCGGGC) [101]

HJ24 Shp2 d(AGCGTCGAATACCACACGGGGGTTTTGGTGGGGGGGGCTGGGTTGTCTTGGGGGTGGGCTAATGGAGCTCGTGGTCAT)

[102]

3R02 VEGF d(TGTGGGGGTGGACTGGGTGGGTACC) [103]

ISIS 5320 HIV gp120 d(T*T*G*G*G*G*T*T) [104]

AS1411 Nucleolin d(GGTGGTGGTGGTTGTGGTGGTGGTGG) [105]

93del HIV integrase d(GGGGTGGGAGGAGGGT) [106]

RT6 HIV reverse transcriptase d(ATCCGCCTGATTAGCGATACTCAGGCGTTAGGGAAGGGCGTCGAAAGCAGGGTGGGACTTGAGCAAAATCACCTGCAGGGG)

[107]

ODN 93 HIV RNase H d(GGGGGTGGGAGGAGGGTAGGCCTTAGGTTTCTGA) [108]

ODN 112 HIV RNase H d(CCAGTGGCGGGTGGGTGGGTGGTGGGGGGACTTGG) [108]

TBA Thrombin d(GGTTGGTGTGGTTGG) [9]

RA-36 Thrombin d(GGTTGGTGTGGTTGGTGGTTGGTGTGGTTGG) [109]

Scl 2 Sclerostin d(TTGCGCGTTAATTGGGGGGGTGGGTGGGTT) [110]

R12 PrPC r(GGAGGAGGAGGA) [111]

PPK2 G9 PPK2 d(AACACATAGGTTTGGTTAGGTTGGTTGGTTGAATTA) [112]

Spinach DFHBI r(GACGCAACUGAAUGAAAUGGUGAAGGACGGGUCCA-GGUGUGGCUGCUUCGGCAGUGCAGCUUGUUGAGUAGA-GUGUGAGCUCCGUAACUAGUCGCGUC)

[14]

Mango TO1 r(UACGAAGGGACGGUGCGGAGAGGAGAGUA) [19]

PS2.M Hemin d(GTGGGTAGGGCGGGTTGG) [23]

* indicates phosphorothioate bond.aAbbreviations: DFHBI, 3,5-difluoro-4-hydroxybenzylidene imidazolinone; PPK2, polyphosphate kinase 2; PrPC, cellular prion protein; STAT3, signal transducer andactivator of transcription 3; TO1, thiazole orange; VEGF, vascular endothelial growth factor.

To address this issue, biochemical techniques were employed to interrogate G-quadruplexformation in a longer sequence context (Table 2). In the DNA polymerase stop assay, theformation of a G-quadruplex in a DNA template can act as a roadblock and cause polymerasestalling, which halts the primer extension. Salazar and colleagues [62] previously applied thismethod to study the DNA G-quadruplex structure formed by telomeric DNA sequences, d(T2G4)4 or d(T2AG4)4, in the template strand. The dimethyl sulfate (DMS) followed by thepiperidine cleavage assay is based on the fact that the formation of a G-quadruplex willprohibit the N7 guanine methylation caused by DMS, leading to a protection pattern observedat the DNA G-quadruplex region after piperidine cleavage. For example, Cech and colleagues[63] used this technique to interrogate telomeric DNA sequences and observed such DMSprotection pattern in the G-quadruplex site. In-line probing is a slow, spontaneous RNAcleavage reaction that measures the flexibility of each RNA nucleotide: this method was firstdeveloped to study the structure of riboswitches, and later applied to RNA G-quadruplexes[64]. Several recent studies have reported the use of in-line probing to probe the formation ofG-quadruplexes in messenger RNAs [48,65].

Recently, several new biochemical methods were developed to study RNA G-quadruplexes(Table 2). Reverse transcriptase can be stalled by RNA G-quadruplex structures during reversetranscription. Kwok and Balasubramanian [66] developed a reverse transcriptase stalling (RTS)assay and coupled this with ligation-mediated PCR to identify the in vitro G-quadruplex

Trends in Biotechnology, October 2017, Vol. 35, No. 10 1001

Figure 2. G-Quadruplexes As Molecular Tools for Biological Applications. (A) Schematic representation of SELEX. Random single-stranded (ss) DNA or RNAoligonucleotides are synthesized, and are subjected to interaction with the target of interest in the selection step. After that, the washing step removes the unboundoligonucleotides and retains the tightly bound ones. These bound oligonucleotides are then eluted out and are amplified by PCR (for ssDNA) or RT-PCR (for ssRNA).After several cycles, the final candidates are cloned and sequenced to identify the DNA or RNA sequence. (B) The G-quadruplex domains of spinach RNA aptamer (PDBID: 4KZD). The figures were adapted and modified with permission from [15]. (C) Schematic representation of hemin–G-quadruplex horseradish peroxidase-mimickingDNAzyme. G-rich sequences such as PS2.M can fold into a G-quadruplex in the presence of K+ ion, which will then bind to hemin. The hemin–quadruplex complexpossesses peroxidase properties that can catalyse the conversion of hydrogen peroxide to water. Colorimetric substrates such as 3,30,5,50-tetramethylbenzidine (TMB)or 2,20-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) can be used to monitor the process. (D) Schematic representation of quadruplex priming amplification.The 2-aminopurine (2AP) containing G-rich primer strand (red) that is missing one track of Gs required to form a G-quadruplex binds with the C-rich template strand(blue). Addition of DNA polymerase allows the missing Gs to be filled in, causes duplex dissociation, and induces G-quadruplex formation. The formation of theG-quadruplex leaves the 2AP to be unstacked at the loop of the G-quadruplex, thus producing fluorescence signal for detection. DFHBI, 3,5-difluoro-4-hydro-xybenzylidene imidazolinone; SELEX, systematic evolution of ligands by exponential enrichment.

formation in low-abundance human telomerase RNA. RNA structure can be probed bychemical probes such as DMS and selective 20-hydroxyl acylation analysed by primerextension (SHAPE) reagents [67,68], and analysed by primer extension. The same groupreported the novel use of lithium-based primer extension (LiPE) buffer in reverse transcriptionthat alleviates RTS, and coupled it with SHAPE reagents and DMS to develop selective 20-hydroxyl acylation analysed by lithium-based primer extension (SHALiPE) and DMSLiPE,respectively [69]. Application of in vitro SHALiPE and DMSLiPE has revealed G-quadruplexformation in precursor microRNA 149 [69]. Dominguez and co-workers [70] developed amethod called ‘footprinting of long 7-deazaguanine-substituted RNAs’ (FOLDeR), which com-pared the RNase footprinting results between wild-type and 7-deazaguanine-substituted RNA.Results from FOLDeR have revealed the in vitro formation and location of RNA G-quadruplexesin a 681-nt fragment of Bcl-x RNA [70].

Most of the biophysical and biochemical assays described here (Table 2) are limited to in vitrostudies; however, several methods can be adapted for in vivo applications, such as the DMS

1002 Trends in Biotechnology, October 2017, Vol. 35, No. 10

Table 2. Representative Biophysical and Biochemical Methods to Study G-Quadruplexes

Methoda G-quadruplex information Features and limitations Refs

CD Topology DNA and RNA, short oligonucleotide, not applicable in vivo [57]

UV melting Thermostability DNA and RNA, short oligonucleotide, not applicable in vivo [58]

FRET melting Thermostability DNA and RNA, short oligonucleotide, not applicable in vivo [113]

NMR G-quartet imino protons at 10.5–12.0 ppm3D structure

DNA and RNA, short oligonucleotide, not directly applicable in vivo [114]

X-ray crystallography 3D structure DNA and RNA, short oligonucleotide, not applicable in vivo [115]

Fluorescent probes(e.g., NMM, ThT)

Presence/absence DNA and RNA, short oligonucleotide, not applicable in vivo [59,60]

TDS Presence/absence DNA and RNA, short oligonucleotide, not applicable in vivo [116]

Optical tweezer Mechanism stability DNA and RNA, short oligonucleotide, not applicable in vivo [117]

Polymerase stop assay Starting location at nucleotide resolution DNA, no information on loops and other G-tracks, not applicable in vivo [62]

DMS and piperidinecleavage assay

Location and structural reactivity at guaninenucleotide resolution

DNA, applicable in vivo, no information on loops [63]

RNase T1 Location and structural reactivity at guaninenucleotide resolution

RNA, no information on loops, not applicable in vivo [118]

In-line probing Location and structural reactivity at singlenucleotide resolution

RNA, react with 20-OH, not applicable in vivo [64]

RTS Starting location at nucleotide resolution RNA, no information on loops and other G-tracks, not applicable in vivo [66]

SHALiPE (and DMSLiPE) Location and structural reactivity at singlenucleotide resolution

RNA, applicable in vivo, react with 20-OH [69]

FOLDeR Location and structural reactivity at singlenucleotide resolution

RNA, requires multiple RNases and 7-deazaguanine substitution, notapplicable in vivo

[70]

aAbbreviations: FRET, fluorescence resonance energy transfer; NMM, N-methyl mesoporphyrin IX; TDS, thermal difference spectra.

and piperidine cleavage assay, SHALiPE, and DMSLiPE. Complementing these excitingbiochemical methods with functional assays (e.g., reporter genes, western blotting, RNAprocessing assays) and cell imaging experiments (as discussed below) will enable us to uncoverthe structural and functional role of G-quadruplexes in cells.

Visualizing G-Quadruplexes Using Cell Imaging MethodsMethods for detecting G-quadruplexes in whole cells have advanced significantly in recent years(Table 3). Two structure-specific antibodies are now available to facilitate G-quadruplex immu-nofluorescence in a rangeof eukaryoticcells [71,72] and in cells infected withG-quadruplexes-richviruses [73]. In parallel, a range of ‘light-up’ chemical probes are rapidly being developed, with thepotential advantage over antibodies that they could be deployable in living cells.

The first quadruplex-specific antibody, Sty49, was used over a decade ago to visualize G-quadruplexes in ciliate macronuclei [74], where large amounts of telomeric DNA offer a super-abundance of G-quadruplexes. A long delay then ensued before successful detection wasreported in fixed mammalian cells (where telomere repeats are much less abundant). Whole-cell immunofluorescence assays may be particularly challenging because chromatin canobscure G-quadruplex epitopes, and because G-quadruplexes may be dynamic and fold onlytransiently in vivo. Several different protein probes were engineered, including a zinc-fingerprotein GQ1 [75], a range of designed ankyrin repeat binding proteins or ‘DARPins’ [76], and asingle-chain antibody hf2 [77]: these could all detect G-quadruplexes in vitro but provedunsuitable for whole-cell immunofluorescence [76]. Since hf2 could be used to pull-down

Trends in Biotechnology, October 2017, Vol. 35, No. 10 1003

Table 3. Representative G-Quadruplex-Specific Antibodies and Chemical Probes for Cell Imaging

Antibodies/probes Application to date Features and limitations Refs Commercially available?

GQ1 zinc-finger protein In vitro detection of DNA G4s Not applied in whole-cell IFA [75] No, see [75] for synthesis

G4 DARPins In vitro detection of DNA G4s Not successful in whole-cell IFA [76] No, see [76] for synthesis

hf2 Single-chain antibody In vitro detection of DNA G4s, pull-down ofG4s from genomic DNA

Not applied in whole-cell IFA [77,78] No, see [77] for synthesis

Sty49 single-chain antibody In vitro detection of DNA G4s, IFA on fixedciliate cells

Detects only high-abundance telomericG4 DNA in macronuclei (fixed cells)

[74] No, see [74] for synthesis

BG4 single-chain antibody IFA on human cells, DNA, and RNA G4s Requires three-step antibody stainingprotocol. Sensitivity to single-G4 levelunproven. Used on fixed cells

[71,79] Yes

1H6 mouse monoclonal antibody IFA on human cells, DNA G4s only Two-step antibody staining protocol.Detects DNA not RNA G4s; cross-reactionrecently reported with poly-T DNA.Sensitivity to single-G4 level unproven.Used on fixed cells.

[72,73,119] Yes

3,6-bis(1-methyl-4-vinylpyridinium)carbazole diiodide

DNA G4s in human cells Light-up, cell permeable (i.e., live cells).Can induce G4s folding in vivo

[82] No, see [120] for synthesis

Squarylium dye TSQ1 DNA G4s in human cells Light-up, cell permeable (i.e., live cells).Does not induce G4s folding in vivo

[83] No, see [83] for synthesis

Cyanine dye CyT RNA G4s in human cells Light-up, cell permeable (i.e., live cells) [84] No, see [84] for synthesis

G-quadruplex-triggered fluorogenichybridization probe, ISCH-nras1

RNA G4 in 50-UTR of NRAS mRNA Light-up, gene-specific. Not cellpermeable, not sensitive to natural lowRNA levels

[85] No, see [85] for synthesis

Anthrathiophenedione dye DNA and RNA G4s in human cells Light-up, cell permeable (i.e., live cells) [86] No, see [86] for synthesis

NaphthoTASQ DNA and RNA G4s in human cells Light-up, affinity triggered by contact withG4s (smart probe). Wavelengthincompatible with standard lightmicroscopy

[88] No, see [121] for synthesis

Triangulenium-derivative DAOTA-M2 DNA and RNA G4s Light-up, cell permeable (i.e., live cells),minimal toxicity

[87] No, see [87] for synthesis

Fluorophore-conjugated RHAU helicasepeptide

DNA G4s in vitro Peptide based. Not yet tested in vivo [89] No, see [89] for synthesis

IFA, immunofluorescence assay; UTR, untranslated region.

1004

Trends

in Biotechnology,

Octob

er 2017,

Vol.

35, No.

10

G-quadruplex DNA fragments from purified genomic DNA [78], the primary problem wasprobably not the sensitivity of these tools, but the in cellulo chromatin context.

Two new antibodies, BG4 [71] and 1H6 [72], have recently proved suitable for whole-cellimmunofluorescence on fixed mammalian cells, although questions remain about whether theycan be sensitive to the level of a single G-quadruplex or can only detect high local densities ofmultiple G-quadruplex motifs. The number of BG4 foci detected in fixed human cells is orders ofmagnitude lower than the number of G-quadruplex-forming sequences predicted throughoutthe genome, but dynamic and transient folding might be expected to limit detection in vivo, andmost antibodies will not detect every possible structural variant of G-quadruplex. Indeed,although the two antibodies were generated via different routes, 1H6 by immunizing mice withstable G-quadruplex structures and BG4 by phage display and in vitro selection on suchstructures, both methodologies are highly dependent upon the exact structures chosen asantigens. Concerning RNA G-quadruplexes, 1H6 apparently does not detect these [72],whereas BG4 does [79].

Turning to the development of small-molecule probes for G-quadruplexes, the status of thisfield has been recently reviewed [61,80,81] and several representative probes are presented inTable 3. In brief, probes for use in living cells should be membrane permeable and minimallycytotoxic, as well as highly selective for G-quadruplex motifs, with strong and specific ‘light-up’(or ‘light-off’) emission versus minimal background. Furthermore, they should not actuallyinduce G-quadruplex formation, but simply detect pre-existing G-quadruplexes – a difficultdistinction to assess. This combination of criteria is very demanding and no perfect probe hasyet been reported. However, there are many promising candidates, including sensors for DNA[82,83], RNA [84,85], and both DNA and RNA G-quadruplexes [86–88], based on a wide varietyof chemistries and possessing combinations of properties (e.g., nontoxicity, cell permeability,G4-induction, and specificity or nonspecificity for particular structural topologies) whoserelative importance must be evaluated for particular applications. Certain probes may betargeted to a particular G-quadruplex-encoding sequence by conjugation with a gene-specificoligonucleotide [85]. Another approach is to attach a fluorophore to a known G-quadruplex-binding protein, thus circumventing the inherent tendency of guanines to quench fluorescencefrom small molecules, and potentially also mitigating cytotoxicity. A G-quadruplex-bindingpeptide from the RHAU helicase has recently showed promise as a sensor in vitro [89].Importantly, all such probes will be subject to the same unanswered question as antibodiesconcerning their sensitivity: can a single G-quadruplex motif ever be detected in cellulo? As anultimate goal for this field, such spatial (and, in live cells, temporal) resolution could beenormously powerful, allowing researchers to determine where and under what cellular con-ditions particular G-quadruplexes can fold, and how they might respond to changing conditionssuch as transcriptional stimuli, DNA damage, or different cell cycle phases.

Genome/Transcriptome-wide Mapping of G-Quadruplexes UsingSequencing MethodsIn contrast to the aforementioned challenge of detecting single G-quadruplex motifs via whole-cell imaging, the advent of NGS has provided an excellent opportunity to design G-quadruplex-specific NGS methods at a genome/transcriptome-wide level. They can be broadly categorizedinto two approaches: the antibody-mediated pull-down approach and the polymerase stallingapproach.

For the antibody-mediated pull-down approach, successful chromatin immunoprecipita-tion (ChIP) of G-quadruplexes was reported only very recently – possibly because suitableantibodies proved elusive, because the native chromatin context tends to mask the majority ofG-quadruplex epitopes, and/or because the PQSs are folded into G-quadruplexes only in

Trends in Biotechnology, October 2017, Vol. 35, No. 10 1005

specific cellular conditions. Circumventing some of these issues, an antibody pull-down wasconducted from naked genomic DNA using the hf2 antibody (Figure 3A), but this detected only�700 G-quadruplexes [78]; subsequently, the BG4 antibody (which was also used in achromatin context for whole-cell immunofluorescence assays) was used to isolateG-quadruplex-containing fragments from human chromatin, yielding �10 000 or 1000 motifsfrom two different human cell lines (Figure 3B) [90]. The isolated regions tended to benoncoding regulatory regions of highly transcribed genes, suggesting that nucleosomedepletion and active transcription probably favour the folding of G-quadruplexes [90]. In a

Figure 3. Schematics of G-Quadruplex-Specific Next-Generation Sequencing (NGS) Methods. (A) hf2 G4 pull-down sequencing [78]: genomic DNAisolated, fragmented by sonication, and incubated with G4-specific hf2 antibody. Mixture washed to remove hf2-unbound DNA before eluting bound DNA. RecoveredDNA library prepared for NGS. (B) BG4 ChIP-seq [3]: cells fixed with formaldehyde, then chromatin isolated, and fragmented by sonication. Fragmented chromatinincubated with G4-specific BG4 antibody with a FLAG-tag. Mixture washed to remove BG4-unbound chromatin, before eluting bound chromatin. Recovered DNAlibrary prepared for NGS. (C) G4-seq [50]: genomic DNA isolated and fragmented by sonication. Fragmented DNA library prepared for NGS. Template DNA is firstsequenced under Na+-containing condition to yield Read 1. The newly synthesized strand is denatured and washed away. The template DNA is renatured with freshprimer and sequenced again under K+-containing (or Na+ + PDS-containing) conditions to yield Read 2. (D) rG4-seq [56]: transcriptomic RNA isolated from culturedcells, polyA-RNA selected, and fragmented by hydrolysis. 30-Adapter ligation performed to provide a handle for reverse transcription in the subsequent step. LigatedRNA divided into three reactions (Li+ containing, K+ containing, and K+ + PDS containing) and renatured, followed by reverse transcription. cDNA fragments ligated to a50-adaptor, followed by PCR and NGS.

1006 Trends in Biotechnology, October 2017, Vol. 35, No. 10

related ‘indirect ChIP’ approach [91], sites of DNA damage marked by the histone variant H2AXwere pulled down after treating cells with the quadruplex-stabilizing drug pyridostatin, on thehypothesis that persistent quadruplexes would induce transcription- and/or replication-dependent DNA damage. This yielded large sequence domains because the histone markspreads broadly at sites of damage, but the domains were indeed enriched in putativequadruplex-forming sequences [91].

Polymerase stalling approaches such as G4-seq offer, by contrast, the comprehensive experi-mental identification of sequences that ‘can’ form G-quadruplexes (Figure 3C) [50], yielding anin vitro genome-wide G-quadruplex map. Here, sheared DNA is subjected to NGS in thepresence or absence of conditions that favour quadruplex folding (potassium ions and/or theG-quadruplex-stabilizing ligand pyridostatin). Under stabilizing conditions, G-quadruplexesimpede the polymerase, causing a characteristic increased mutation rate in sequence dataat the G-quadruplex folded region. The G4-seq technique identified �700 000 G-quadruplexesin the human genome: orders of magnitude more than ChIP and two times the numberpredicted in silico by standard algorithms. This is because many of the sequenced motifswere bulged or long looped – and indeed, the majority of G-quadruplexes found by ChIP (79%)were not of the canonical G3 N1–7-type either [52]. However, those that were of this type tendedto represent the strongest ChIP peaks, and there was also a reassuring degree of consonancebetween the ChIP and G4-seq experiments: 87% of the DNA fragments from ChIP containedsequences identified in G4-seq [50].

The difference in the number of G4s reported from data sets obtained in vitro and in vivo(i.e., G4-seq vs. G4-ChIP) suggests that the cellular environment may play a central role inaffecting the dynamics of G-quadruplex formation in cells. Specifically, many cellular factors,such as G-quadruplex binding proteins and helicases (Box 1), can likely remodel the DNAG-quadruplex landscape in vivo. Nevertheless, more technical explanations for the difference inG-quadruplex detection rates in vivo versus in vitro cannot be excluded, such as limitedsensitivity of ChIP and condition dependence for G-quadruplex formation in native chromatin.Results are also likely to be influenced by sensitivity/specificity of the antibody used, choice ofcell line, and variation in experimental protocols and bioinformatics pipelines. Future develop-ment in G-quadruplex-antibody-based sequencing methods should aim to address theseissues to establish a gold standard for robust mapping of DNA G-quadruplexes in nativechromatin for different species.

Turning from DNA to RNA G-quadruplexes, a polymerase stalling approach called rG4-seq wasrecently reported (Figure 3D) [56], based on the working principle first developed in RTS(described above) [66]. Here, RNA G-quadruplexes impede the reverse transcriptase enzymeused to generate an NGS library from polyA-enriched RNAs. rG4-seq is particularly useful forgenerating an in vitro transcriptome-wide map of RNA G-quadruplexes, which will serve as auseful guide for future in vivo studies. Application of rG4-seq has identified a preponderance ofunconventional G-quadruplexes such as long loops, bulged, and two-quartet structures.Notably, significant correlations were reported between RNA G-quadruplexes and cis-regula-tory elements such as microRNA target sites and polyadenylation signals. Further in-depthstudies may examine how RNA G-quadruplexes may regulate microRNA targeting and alter-native polyadenylation to affect gene expression and RNA metabolism in cells.

As discussed earlier and shown in Table 2 (see DMSLiPE and SHALiPE), chemical-based RNAstructure probing assays [69] have the advantages to probe in vivo RNA structure on atranscriptome-wide scale [67,68,92]. A recent report by Guo and Bartel [93] used DMSand SHAPE chemical probing and reported that RNA G-quadruplexes are under-representedin bacteria, and are globally unfolded in human, mouse, and yeast, suggesting that eukaryotic

Trends in Biotechnology, October 2017, Vol. 35, No. 10 1007

Box 1. Proteins That Bind and/or Metabolize G-Quadruplexes

Many cellular proteins have been identified that interact with DNA and/or RNA quadruplexes: many of these are listed inthe G4IPDB database (http://bsbe.iiti.ac.in/bsbe/ipdb/index.php) and some of the key players are tabulated in Table I.These proteins can be used as tools to probe the distribution and function of quadruplex motifs, as well as beingsubjects of intense study themselves.

DNA G-Quadruplexes

Proteins partners of DNA G-quadruplexes include several groups of structure-specific helicases, such as PIF1, RECQ,and FANCJ (recently reviewed in [122]); the transcriptional helicase complementation group XPD/XPB [123]; andvarious nonhelicase proteins including nucleolin [124], shelterin components like POT1 [125], and transcription factorslike Sp1 and MAZ [126,127]. Deficiencies in RECQ, FANCJ, and XP helicases are linked to rare human diseasesyndromes: Fanconi’s anaemia for FANCJ, xeroderma pigmentosum for XPD/XPB and Bloom’s, Werner’s, andRothmund–Thomson syndromes for three members of the five-member RECQ family. The diseases are generallycharacterized by chromosomal instability, telomere deficiency, cancer proneness, etc.: the expected phenotypes forcells that cannot resolve noncanonical DNA secondary structures such as G-quadruplexes, and hence suffer high ratesof DNA replication fork stalling. At a molecular level, ChIP has demonstrated that these helicases tend to associate withPQSs in the genome, particularly when cells are treated with G-quadruplex-stabilizing drugs [128], while in helicase-deficient cells, genes whose expression is deregulated likewise tend to contain PQSs [129]. However, the correlation isnot direct and exclusive because some of the helicases also target other structures such as hairpins and chicken-footstructures [130]. Accordingly, ChIP experiments for PIF1 and RECQs may yield many more targets than direct G4-ChIP.

RNA G-Quadruplexes

Protein partners of RNA G-quadruplexes include helicases such as RHAU (DHX36) DHX9, as well as nonhelicaseproteins such as FMRP and Aven. For more details, please see recent excellent reviews [4,100,131]. RHAU is one of themost studied helicases for RNA G-quadruplexes. RHAU is shown to be involved in the maturation of human telomeraseRNA (hTERC) by unwinding the RNA G-quadruplex at the 50-end of hTERC [132]. FMRP is an important protein that isresponsible for fragile X syndrome, and is crucial for cognitive development. Binding assays and bioinformatics analysisof NGS data suggested that it interacts with RNA G-quadruplexes [133,134]. Recently, a crystal structure revealed thatit requires an RNA duplex–quadruplex junction for recognition [135]. Similar to DNA G-quadruplex binding proteins,RNA G-quadruplex binding proteins also target other structures such as triple helices [134,136,137]. Thus, one shouldbe cautious about the RNA–protein NGS data obtained on these RNA G-quadruplex binding proteins, as they likely alsocontain structural motifs that do not fold into G-quadruplexes.

Table I. Representative G-Quadruplex-Interacting Proteins

Protein name/family DNA or RNAquadruplexes

Role Refs

PIF1 DNA Structure-specific 50–30 helicase, interacts with telomerase,and regulates telomere maintenance

[138]

RECQ family DNA Structure-specific 30-50 helicases, act at noncanonical DNAstructures, facilitate DNA repair, and suppressrecombination

[122]

BRIP1 (FANCJ)a DNA Structure-specific 50–30 helicase (RAD 3 family), primarilyinvolved in repair of DNA crosslinks

[122]

ERCC2/3 (XPB/XPD)a DNA Helicase subunits of the TFIIH transcription/repair factor,involved in nucleotide excision repair

[123]

Nucleolin DNA Nucleolar protein, controls rRNA gene transcription, andassembly/export of ribosomes

[124]

POT1 DNA Shelterin component, protects telomeres [125]

PARP-1 DNA Poly(ADP-ribose)polymerase, ADP ribosylates manychromatin proteins and is involved in DNA repair

[139]

DHX36 (RHAU)a RNA and DNA RNA (and DNA) helicase with preference for RNA G-quadruplexes

[132]

DHX9 RNA RNA helicase with preference for RNA G-quadruplexes [140]

FMRP RNA Binds to mRNAs and regulates association with polysomes [133,134]

1008 Trends in Biotechnology, October 2017, Vol. 35, No. 10

Table I. (continued)

Protein name/family DNA or RNAquadruplexes

Role Refs

Aven RNA Regulator of apoptosis. Also associates with DHX36helicase and binds mRNA G-quadruplexes

[141]

HnRNP F RNA Pre-mRNA processing and translocation [142]

DDX21 RNA RNA helicase with affinity for RNA G-quadruplexes [143]

CNBP RNA Zinc finger protein controlling the translational efficiency ofmRNAs

[144]

EBNA1 RNA Viral protein involved in controlling replication of Epstein–Barr virus

[145]

aGiven in parentheses are alternative names or the names of disease complementation groups caused by mutations ofthe given genes.

cellular machinery may actively control or unwind RNA G-quadruplexes. These results advanceour understanding of the dynamic (un)formation and role of RNA G-quadruplexes in cells, asmany G-quadruplexes were reported to form in vitro [56,93], yet less so in vivo [93]. Futureexperiments may investigate whether these phenomena are generally applicable to other celltypes, cellular conditions, and species by using multiple RNA structurome and interactomemethods [67,68,92], and orthogonal G-quadruplex antibody-mediated pull-down approachesas similarly performed for DNA G-quadruplexes. One future challenge is to identify andcharacterize the effect of known and unknown RNA G-quadruplex binding proteins (Box 1)on G-quadruplex structure and function. Overall, these genome- and transcriptome-widestudies generate new testable hypotheses and offer future directions to explore the G-quad-ruplex-mediated biological processes across the tree of life.

Future Perspectives and ChallengesAmong the technologies discussed in this review, some are fairly advanced, while others remainin their infancy. Biophysical – and to some extent, biochemical – methods for identifyingG-quadruplexes have been developed and applied for many years and a wealth of in silicoalgorithms is available for predicting G-quadruplex formation from nucleic acid sequences.However, it is increasingly clear that the rules for G-quadruplex folding are complex, subtle, andcontext dependent. Few of the algorithms comprehensively incorporate empirical experimentaldata, few biophysical techniques incorporate the wider sequence context, and few biochemicaltechniques are applicable in vivo to date (Table 2). As such, the field calls for new in vivomethods with superior resolution, throughput, and sensitivity to investigate the spatial–tempo-ral formation of G-quadruplexes, their structure folding and dynamics, and the effect of cellularfactors upon G-quadruplexes that prevail in cellular milieu.

Cell imaging methods have advanced tremendously in the past few years, but challengesremain around the feasibility of resolving single G-quadruplex motifs, the relative accessibility ofG-quadruplexes in chromatin contexts, and the potentially transient nature of many G-quad-ruplexes in vivo. In addition, it is yet to be seen if the G-quadruplex-specific probes (Table 3) canbe easily applied to other biological systems. Notably, the cellular localization and live cellimaging of G-quadruplexes in particular genes are largely untested: this would requiresequence-specific G-quadruplex antibodies, oligonucleotide-conjugated G-quadruplexprobes, or a combination of immunofluorescence assay and fluorescent in situ hybridization(both of which are highly demanding in terms of sensitivity). In addition, the production of theperfect light-up G-quadruplex probes for use in living cells remains elusive, although somerecent progress has been made toward this goal [61,80,81].

Trends in Biotechnology, October 2017, Vol. 35, No. 10 1009

Outstanding QuestionsDynamic regulation of G-quadruplexesin vivo: Under what conditions and inwhich cellular compartments do theyform? How are they regulated?

Cellular imaging: Can G-quadruplexesbe imaged in living, as well as fixed,cells in different biological systems?Can they be imaged at single-motifresolution?

In vivo G-quadruplex ‘omics’: Can invivo genome-wide and transcriptome-wide mapping of G-quadruplex struc-tures and G-quadruplex protein inter-actions be achieved in diverseorganisms? Can new sequencingmethods be developed to complementthe current approaches?

Intermolecular G-quadruplexes: Canintermolecular G-quadruplex motifsinvolving DNA, RNA, or DNA–RNAhybrids be identified and characterizedin vivo? Can new chemicals andmolecular tools be developed toachieve this?

G-quadruplexes across the tree of life:how are G-quadruplexes used in thebiology of diverse species of prokar-yotes, eukaryotes, and viruses? Arethey evolutionarily conserved or spe-cies specific?

The link between G-quadruplexes anddiseases: What are the underlying bio-chemical mechanisms? Can new G-quadruplex-based diagnostic andtherapeutic applications bedeveloped?

Like whole-cell imaging, the ‘omic’-level sequencing techniques now available have explodedrecently with a series of seminal papers. In this field, defining the formation, structure dynamics,and interaction partners of G-quadruplexes in cellulo remains the key question. G-quadruplexformation can be influenced by a variety of factors, including metal ions, flanking sequencecontext, and proteins. Recent studies showed that G-quadruplexes could interconvert withstem–loop structures to regulate cellular processes [69,94–96], suggesting that alternativestructures such as duplexes or hairpins may compete with G-quadruplex formation in vivo. Itwill be interesting to structurally probe them in vivo and see how prevalent these G-quadruplexstructure switches are, and what are their regulatory roles in vivo. As mentioned earlier anddetailed in Box 1, some known G-quadruplex binding proteins and helicases have beenidentified; however, their global effects on G-quadruplex structures in cells are largely unchar-acterized. In the future, further studies performed under knock-down/knockout of G-quad-ruplex binding proteins, and under normal and stress conditions, could provide clues about theformation, structure dynamics, and interactions of G-quadruplexes in vivo.

Most G-quadruplex studies consider only intramolecular G-quadruplex folding; however,bioinformatics searches have shown the potential prevalence of intermolecular DNA–RNAG-quadruplexes in humans [97]. Given the vast number of predicted intermolecular G-quad-ruplexes, great experimental effort and robust analysis platforms are needed to reveal theirpervasiveness, their structural conformational exchange with intramolecular G-quadruplexes orother structural motifs, and their potential functions in cells, such as in transcription. Innovativestrategies are thus urgently needed to be able to detect and characterize these intermolecularG-quadruplex motifs in vivo.

Concluding RemarksRemarkable progress has been made in G-quadruplex research in the past 5 years. This is anexciting time to explore the in vivo G-quadruplex structure at unprecedented resolution, through-put, and sensitivity. We are cautiously positive that the development of a suite of novel methodol-ogies will reveal the in vivo structures and functions of G-quadruplexes in diverse organisms, andhelp to address the biological questions regarding the prevalence, location, diversity, dynamics,interactions, and localization of G-quadruplexes in diverse organisms (see OutstandingQuestions). As emerging evidence suggests a connection between G-quadruplexes, generegulation, and development of diseases [98–100], these upcoming advancements in G-quad-ruplex research will decipher the underlying biochemical mechanism and the molecular basis ofdiseases, and also facilitate the rational design and development of G-quadruplex-related tools forvarious biological applications. We look forward with great optimism to the next set of ground-breaking discoveries and applications to be unveiled in the next 5 years.

AcknowledgementsThis work is supported by grants from City University of Hong Kong (Project No. 9610363, 7200520) and Croucher

Foundation (Project No. 9500030) to C.K.K., and grants from the UK Medical Research Council (MR/L008823/1 and MR/

P010873/1) to C.J.M. We acknowledge S.M. Leung for helping with figures. We apologize to colleagues whose works are

not cited due to space limitation.

References

1. Bloomfield, V.A. et al. (2000) Nucleic Acids: Structures, Proper-

ties, and Functions, Sterling Publishing Company

2. Neidle, S. and Balasubramanian, S. (2006) Quadruplex NucleicAcids, Royal Society of Chemistry

3. Hansel-Hertsch, R. et al. (2017) DNA G-quadruplexes in thehuman genome: detection, functions and therapeutic potential.Nat. Rev. Mol. Cell Biol. 18, 279–284

4. Millevoi, S. et al. (2012) G-quadruplexes in RNA biology. WileyInterdiscip. Rev. RNA 3, 495–507

1010 Trends in Biotechnology, October 2017, Vol. 35, No. 10

5. Collie, G.W. and Parkinson, G.N. (2011) The application of DNAand RNA G-quadruplexes to therapeutic medicines. Chem.Soc. Rev. 40, 5867–5892

6. Tian, T. et al. (2015) G-quadruplex’s applications in biologicaltarget detection and drug delivery. Curr. Top. Med. Chem. 15,1988–2001

7. Platella, C. et al. (2016) G-quadruplex-based aptamers againstprotein targets in therapy and diagnostics. Biochim. Biophys.Acta 1861, 1429–1447

8. Viglasky, V. and Hianik, T. (2013) Potential uses of G-quadru-plex-forming aptamers. Gen. Physiol. Biophys. 32, 149–172

9. Bock, L.C. et al. (1992) Selection of single-stranded-DNAmolecules that bind and inhibit human thrombin. Nature 355,564–566

10. Russo Krauss, I. et al. (2012) High-resolution structures of twocomplexes between thrombin and thrombin-binding aptamershed light on the role of cations in the aptamer inhibitory activity.Nucleic Acids Res. 40, 8119–8128

11. Nagatoishi, S. et al. (2007) Circular dichroism spectra demon-strate formation of the thrombin-binding DNA aptamer G-quad-ruplex under stabilizing-cation-deficient conditions. Biochem.Biophys. Res. Commun. 352, 812–817

12. Deng, B. et al. (2014) Aptamer binding assays for proteins: thethrombin example – a review. Anal. Chim. Acta 837, 1–15

13. Wang, Q. et al. (2015) Label-free aptamer biosensor for throm-bin detection based on functionalized graphene nanocompo-sites. Talanta 141, 247–252

14. Paige, J.S. et al. (2011) RNA mimics of green fluorescent pro-tein. Science 333, 642–646

15. Huang, H. et al. (2014) A G-quadruplex-containing RNA acti-vates fluorescence in a GFP-like fluorophore. Nat. Chem. Biol.10, 686–691

16. Strack, R.L. et al. (2014) Using spinach-based sensors forfluorescence imaging of intracellular metabolites and proteinsin living bacteria. Nat. Protoc. 9, 146–155

17. Strack, R.L. and Jaffrey, S.R. (2015) Live-cell imaging of mam-malian RNAs with spinach2. Methods Enzymol. 550, 129–146

18. Autour, A. et al. (2016) iSpinach: a fluorogenic RNA aptameroptimized for in vitro applications. Nucleic Acids Res. 44, 2491–2500

19. Dolgosheina, E.V. et al. (2014) RNA mango aptamer-fluoro-phore: a bright, high-affinity complex for RNA labeling andtracking. ACS Chem. Biol. 9, 2412–2420

20. Okuda, M. et al. (2016) Use of baby spinach and broccoli forimaging of structured cellular RNAs. Nucleic Acids Res. 45,1404–1415

21. Jeng, S.C. et al. (2016) Fluorophore ligand binding and complexstabilization of the RNA mango and RNA spinach aptamers.RNA 22, 1884–1892

22. Trachman, R.J., 3rd et al. (2017) Structural basis for high-affinityfluorophore binding and activation by RNA mango. Nat. Chem.Biol. 13, 807–813

23. Travascio, P. et al. (1998) DNA-enhanced peroxidase activity ofa DNA-aptamer-hemin complex. Chem. Biol. 5, 505–517

24. He, Y. et al. (2009) Human telomeric DNA sequences have aperoxidase apoenzyme activity. Mol. Biosyst. 5, 238–240

25. Kong, D.M. et al. (2009) Characterization of the G-quadruplexstructure of a catalytic DNA with peroxidase activity. Biopoly-mers 91, 331–339

26. Kong, D.M. et al. (2010) Structure–function study of peroxidase-like G-quadruplex-hemin complexes. Analyst 135, 321–326

27. Poon, L.C.H. et al. (2011) Guanine-rich RNAs and DNAs thatbind heme robustly catalyze oxygen transfer reactions. J. Am.Chem. Soc. 133, 1877–1884

28. Loic, S. (2016) Native and synthetic G-quartet-based DNAzymesystems – artificial enzymes for biotechnological applications. InAdvanced Catalytic Materials – Photocatalysis and OtherCurrent Trends (Norena, L.E. and Wang, J.-A., eds),pp. 447–485, InTech

29. Taylor, A. et al. (2013) Isothermal quadruplex priming amplifica-tion for DNA-based diagnostics. Biophys. Chem. 171, 1–8

30. Kankia, B.I. (2011) Self-dissociative primers for nucleic acidamplification and detection based on DNA quadruplexes withintrinsic fluorescence. Anal. Biochem. 409, 59–65

31. Van Ness, J. et al. (2003) Isothermal reactions for the amplifica-tion of oligonucleotides. Proc. Natl. Acad. Sci. U. S. A. 100,4504–4509

32. Gogichaishvili, S. et al. (2014) Quadruplex priming amplificationcombined with nicking enzyme for diagnostics. Anal. Biochem.466, 44–48

33. Mendez, M.A. and Szalai, V.A. (2009) Fluorescence of unmodi-fied oligonucleotides: a tool to probe G-quadruplex DNAstructure. Biopolymers 91, 841–850

34. Dao, N.T. et al. (2011) Following G-quadruplex formation by itsintrinsic fluorescence. FEBS Lett. 585, 3969–3977

35. Kwok, C.K. et al. (2013) Decrease in RNA folding cooperativityby deliberate population of intermediates in RNA G-quadru-plexes. Angew. Chem. Int. Ed. 52, 683–686

36. Kwok, C.K. et al. (2013) Effect of loop sequence and loop lengthon the intrinsic fluorescence of G-quadruplexes. Biochemistry52, 3019–3021

37. Dao, N.T. et al. (2013) Excimer formation by stacking G-quad-ruplex blocks. Chemphyschem 14, 2667–2671

38. Sherlock, M.E. et al. (2016) Steady-state and time-resolvedstudies into the origin of the intrinsic fluorescence of G-quadruplexes. J. Phys. Chem. B 120, 5146–5158

39. Huppert, J.L. and Balasubramanian, S. (2005) Prevalence ofquadruplexes in the human genome. Nucleic Acids Res. 33,2908–2916

40. Kikin, O. et al. (2006) QGRS Mapper: a web-based server forpredicting G-quadruplexes in nucleotide sequences. NucleicAcids Res. 34, W676–W682

41. Eddy, J. and Maizels, N. (2006) Gene function correlates withpotential for G4 DNA formation in the human genome. NucleicAcids Res. 34, 3887–3896

42. Yadav, V.K. et al. (2008) QuadBase: genome-wide database ofG4 DNA – occurrence and conservation in human, chimpanzee,mouse and rat promoters and 146 microbes. Nucleic Acids Res.36, D381–D385

43. Bedrat, A. et al. (2016) Re-evaluation of G-quadruplex propen-sity with G4Hunter. Nucleic Acids Res. 44, 1746–1759

44. Wong, H.M. et al. (2010) A toolbox for predicting G-quadruplexformation and stability. J. Nucleic Acids 2010, 564946

45. Cao, K. et al. (2012) Computational detection and analysis ofsequences with duplex-derived interstrand G-quadruplex form-ing potential. Methods 57, 3–10

46. Kudlicki, A.S. (2016) G-quadruplexes involving both strands ofgenomic DNA are highly abundant and colocalize with functionalsites in the human genome. PLoS One 11, e0146174

47. Lorenz, R. et al. (2012) RNA folding algorithms with G-quad-ruplexes. In Advances in Bioinformatics and ComputationalBiology (de Souto, M.C.P. and Kann, M.G., eds), pp. 49–60,Springer Berlin Heidelberg

48. Beaudoin, J.D. et al. (2014) New scoring system to identify RNAG-quadruplex folding. Nucleic Acids Res. 42, 1209–1223

49. Stanton, A. et al. (2016) Recombination events among virulencegenes in malaria parasites are associated with G-quadruplex-forming DNA motifs. BMC Genomics 17, 859

50. Chambers, V.S. et al. (2015) High-throughput sequencing ofDNA G-quadruplex structures in the human genome. Nat. Bio-technol. 33, 877–881

51. Guedin, A. et al. (2010) How long is too long? Effects of loop sizeon G-quadruplex stability. Nucleic Acids Res. 38, 7858–7868

52. Piazza, A. et al. (2015) Short loop length and high thermalstability determine genomic instability induced by G-quadru-plex-forming minisatellites. EMBO J. 34, 1718–1734

53. Kim, M. et al. (2016) Quantitative analysis and prediction of G-quadruplex forming sequences in double-stranded DNA.Nucleic Acids Res. 44, 4807–4817

54. Yano, M. and Kato, Y. (2014) Using hidden Markov models toinvestigate G-quadruplex motifs in genomic sequences. BMCGenomics 15, S15

55. Mukundan, V.T. and Phan, A.T. (2013) Bulges in G-quadru-plexes: broadening the definition of G-quadruplex-formingsequences. J. Am. Chem. Soc. 135, 5017–5028

56. Kwok, C.K. et al. (2016) rG4-seq reveals widespread formationof G-quadruplex structures in the human transcriptome. Nat.Methods 13, 841–844

57. Paramasivan, S. et al. (2007) Circular dichroism of quadruplexDNAs: applications to structure, cation effects and ligand bind-ing. Methods 43, 324–331

Trends in Biotechnology, October 2017, Vol. 35, No. 10 1011

58. Mergny, J.L. et al. (1998) Following G-quartet formation by UV-spectroscopy. FEBS Lett. 435, 74–78

59. Renaud de la Faverie, A. et al. (2014) Thioflavin T as a fluores-cence light-up probe for G4 formation. Nucleic Acids Res. 42,e65

60. Kreig, A. et al. (2015) G-quadruplex formation in double strandDNA probed by NMM and CV fluorescence. Nucleic Acids Res.43, 7961–7970

61. Ma, D.L. et al. (2015) Recent developments in G-quadruplexprobes. Chem. Biol. 22, 812–828

62. Han, H. et al. (1999) A DNA polymerase stop assay for G-quadruplex-interactive compounds. Nucleic Acids Res. 27,537–542

63. Williamson, J.R. et al. (1989) Monovalent cation-inducedstructure of telomeric DNA: the G-quartet model. Cell 59,871–880

64. Beaudoin, J.D. et al. (2013) In-line probing of RNA G-quadru-plexes. Methods 64, 79–87

65. Kwok, C.K. et al. (2015) A stable RNA G-quadruplex within the50-UTR of Arabidopsis thaliana ATR mRNA inhibits translation.Biochem. J. 467, 91–102

66. Kwok, C.K. and Balasubramanian, S. (2015) Targeted detectionof G-quadruplexes in cellular RNAs. Angew. Chem. Int. Ed. 54,6751–6754

67. Kwok, C.K. et al. (2015) The RNA structurome: transcriptome-wide structure probing with next-generation sequencing.Trends Biochem. Sci. 40, 221–232

68. Kwok, C.K. (2016) Dawn of the in vivo RNA structurome andinteractome. Biochem. Soc. Trans. 44, 1395–1410

69. Kwok, C.K. et al. (2016) Structural analysis using SHALiPE toreveal RNA G-quadruplex formation in human precursor micro-RNA. Angew. Chem. Int. Ed. 55, 8958–8961

70. Weldon, C. et al. (2017) Identification of G-quadruplexes in longfunctional RNAs using 7-deazaguanine RNA. Nat. Chem. Biol.13, 18–20

71. Biffi, G. et al. (2013) Quantitative visualization of DNA G-quad-ruplex structures in human cells. Nat. Chem. 5, 182–186

72. Henderson, A. et al. (2014) Detection of G-quadruplex DNA inmammalian cells. Nucleic Acids Res. 42, 860–869

73. Artusi, S. et al. (2016) Visualization of DNA G-quadruplexes inherpes simplex virus 1-infected cells. Nucleic Acids Res. 44,10343–10353

74. Schaffitzel, C. et al. (2001) In vitro generated antibodies specificfor telomeric guanine-quadruplex DNA react with Stylonychialemnae macronuclei. Proc. Natl. Acad. Sci. U. S. A. 98, 8572–8577

75. Isalan, M. et al. (2001) Selection of zinc fingers that bind single-stranded telomeric DNA in the G-quadruplex conformation.Biochemistry 40, 830–836

76. Scholz, O. et al. (2014) G-quadruplexes are specifically recog-nized and distinguished by selected designed ankyrin repeatproteins. Nucleic Acids Res. 42, 9182–9194

77. Fernando, H. et al. (2008) Selective recognition of a DNAG-quadruplex by an engineered antibody. Biochemistry 47,9365–9371

78. Lam, E.Y. et al. (2013) G-quadruplex structures are stable anddetectable in human genomic DNA. Nat. Commun. 4, 1796

79. Biffi, G. et al. (2014) Visualization and selective chemical target-ing of RNA G-quadruplex structures in the cytoplasm of humancells. Nat. Chem. 6, 75–80

80. Vummidi, B.R. et al. (2013) Fluorescent probes for G-quadru-plex structures. Chembiochem 14, 540–558

81. Ma, D.L. et al. (2015) Recent development of G-quadruplexprobes for cellular imaging. Curr. Top. Med. Chem. 15,1957–1963

82. Tseng, T.Y. et al. (2013) In-cell optical imaging of exogenous G-quadruplex DNA by fluorogenic ligands. Nucleic Acids Res. 41,10605–10618

83. Chen, Y. et al. (2014) Nonlinear optical dye TSQ1 as an efficientlyselective fluorescent probe for G-quadruplex DNA. Org. Chem.Front. 1, 267–270

1012 Trends in Biotechnology, October 2017, Vol. 35, No. 10

84. Xu, S. et al. (2015) Directly lighting up RNA G-quadruplexes fromtest tubes to living human cells. Nucleic Acids Res. 43, 9575–9586

85. Chen, S.B. et al. (2016) Visualization of NRAS RNA G-quad-ruplex structures in cells with an engineered fluorogenic hybrid-ization probe. J. Am. Chem. Soc. 138, 10382–10385

86. Cogoi, S. et al. (2013) Guanidino anthrathiophenediones asG-quadruplex binders: uptake, intracellular localization, andanti-Harvey-Ras gene activity in bladder cancer cells. J. Med.Chem. 56, 2764–2778

87. Shivalingam, A. et al. (2015) The interactions between a smallmolecule and G-quadruplexes are visualized by fluorescencelifetime imaging microscopy. Nat. Commun. 6, 8178

88. Laguerre, A. et al. (2016) Direct visualization of both DNA andRNA quadruplexes in human cells via an uncommon spectro-scopic method. Sci. Rep. 6, 32141

89. Dang, D.T. and Phan, A.T. (2016) Development of fluorescentprotein probes specific for parallel DNA and RNA G-quadru-plexes. Chembiochem 17, 42–45

90. Hansel-Hertsch, R. et al. (2016) G-quadruplex structures markhuman regulatory chromatin. Nat. Genet. 48, 1267–1272

91. Rodriguez, R. et al. (2012) Small-molecule-induced DNA dam-age identifies alternative DNA structures in human genes. Nat.Chem. Biol. 8, 301–310

92. Lu, Z. and Chang, H.Y. (2016) Decoding the RNA structurome.Curr. Opin. Struct. Biol. 36, 142–148

93. Guo, J.U. and Bartel, D.P. (2016) RNA G-quadruplexes areglobally unfolded in eukaryotic cells and depleted in bacteria.Science 353, aaf5371

94. Pandey, S. et al. (2015) The RNA stem-loop to G-quadruplexequilibrium controls mature microRNA production inside thecell. Biochemistry 54, 7067–7078

95. Kuo, M.H. et al. (2015) Conformational transition of a hairpinstructure to G-quadruplex within the WNT1 gene promoter. J.Am. Chem. Soc. 137, 210–218

96. Onel, B. et al. (2016) A new G-quadruplex with hairpin loopimmediately upstream of the human BCL2 P1 promoter mod-ulates transcription. J. Am. Chem. Soc. 138, 2563–2570

97. Zheng, K.W. et al. (2013) Co-transcriptional formation of DNA:RNA hybrid G-quadruplex and potential function as constitu-tional cis element for transcription control. Nucleic Acids Res.41, 5533–5541

98. Maizels, N. (2015) G4-associated human diseases. EMBO Rep.16, 910–922

99. Simone, R. et al. (2015) G-quadruplexes: emerging roles inneurodegenerative diseases and the non-coding transcriptome.FEBS Lett. 589, 1653–1668

100. Cammas, A. and Millevoi, S. (2016) RNA G-quadruplexes:emerging mechanisms in disease. Nucleic Acids Res. 45,1584–1595

101. Jing, N.J. et al. (2004) G-quartet oligonucleotides: a new class ofsignal transducer and activator of transcription 3 inhibitors thatsuppresses growth of prostate and breast tumors throughinduction of apoptosis. Cancer Res. 64, 6603–6609

102. Hu, J. et al. (2011) A G-quadruplex aptamer inhibits the phos-phatase activity of oncogenic protein Shp2 in vitro. Chembio-chem 12, 424–430

103. Nonaka, Y. et al. (2013) Affinity improvement of a VEGF aptamerby in silico maturation for a sensitive VEGF-detection system.Anal. Chem. 85, 1132–1137

104. Wyatt, J.R. et al. (1994) Combinatorially selected guanosine-quartet structure is a potent inhibitor of human-immunodefi-ciency-virus envelope-mediated cell-fusion. Proc. Natl. Acad.Sci. U. S. A. 91, 1356–1360

105. Bates, P.J. et al. (2009) Discovery and development of theG-rich oligonucleotide AS1411 as a novel treatment for cancer.Exp. Mol. Pathol. 86, 151–164

106. de Soultrait, V.R. et al. (2002) DNA aptamers derived from HIV-1RNase H inhibitors are strong anti-integrase agents. J. Mol. Biol.324, 195–203

107. Michalowski, D. et al. (2008) Novel bimodular DNA aptamerswith guanosine quadruplexes inhibit phylogenetically diverse

HIV-1 reverse transcriptases. Nucleic Acids Res. 36, 7124–7135

108. Andreola, M.L. et al. (2001) DNA aptamers selected against theHIV-1 RNase H display in vitro antiviral activity. Biochemistry 40,10087–10094

109. Zavyalova, E. et al. (2012) DNA aptamers for human thrombinwith high anticoagulant activity demonstrate target- and spe-cies-specificity. Curr. Med. Chem. 19, 5232–5237

110. Shum, K.T. et al. (2011) Identification of a DNA aptamer thatinhibits sclerostin’s antagonistic effect on Wnt signalling. Bio-chem. J. 434, 493–501

111. Mashima, T. et al. (2013) Anti-prion activity of an RNA aptamerand its structural basis. Nucleic Acids Res. 41, 1355–1362

112. Shum, K.T. et al. (2011) Aptamer-mediated inhibition of Myco-bacterium tuberculosis polyphosphate kinase 2. Biochemistry50, 3261–3271

113. Mergny, J.L. et al. (2001) Telomerase inhibitors based on quad-ruplex ligands selected by a fluorescence assay. Proc. Natl.Acad. Sci. U. S. A. 98, 3062–3067

114. Webba da Silva, M. (2007) NMR methods for studying quad-ruplex nucleic acids. Methods 43, 264–277

115. Parkinson, G.N. et al. (2002) Crystal structure of parallel quad-ruplexes from human telomeric DNA. Nature 417, 876–880

116. Mergny, J.L. et al. (2005) Thermal difference spectra: a specificsignature for nucleic acid structures. Nucleic Acids Res. 33,e138

117. Yu, Z. et al. (2009) ILPR G-quadruplexes formed in secondsdemonstrate high mechanical stabilities. J. Am. Chem. Soc.131, 1876–1882

118. Morris, M.J. and Basu, S. (2009) An unusually stable G-quad-ruplex within the 50-UTR of the MT3 matrix metalloproteinasemRNA represses translation in eukaryotic cells. Biochemistry48, 5313–5319

119. Kazemier, H.G. et al. (2017) Guanine quadruplex monoclonalantibody 1H6 cross-reacts with restrained thymidine-rich singlestranded DNA. Nucleic Acids Res. 45, 5913–5919

120. Chang, C.C. et al. (2003) A carbazole derivative synthesis forstabilizing the quadruplex structure. J. Chin. Chem. Soc. 50,185–188

121. Laguerre, A. et al. (2015) Visualization of RNA-quadruplexes inlive cells. J. Am. Chem. Soc. 137, 8521–8525

122. Mendoza, O. et al. (2016) G-quadruplexes and helicases.Nucleic Acids Res. 44, 1989–2006

123. Gray, L.T. et al. (2014) G quadruplexes are genomewide targetsof transcriptional helicases XPB and XPD. Nat. Chem. Biol. 10,313–318

124. Gonzalez, V. et al. (2009) Identification and characterization ofnucleolin as a c-myc G-quadruplex-binding protein. J. Biol.Chem. 284, 23622–23635

125. Ray, S. et al. (2014) G-quadruplex formation in telomeres enhan-ces POT1/TPP1 protection against RPA binding. Proc. Natl.Acad. Sci. U. S. A. 111, 2990–2995

126. Piekna-Przybylska, D. et al. (2014) U3 region in the HIV-1genome adopts a G-quadruplex structure in its RNA andDNA sequence. Biochemistry 53, 2581–2593

127. Cogoi, S. et al. (2014) HRAS is silenced by two neighboring G-quadruplexes and activated by MAZ, a zinc-finger transcriptionfactor with DNA unfolding property. Nucleic Acids Res. 42,8379–8388

128. Paeschke, K. et al. (2011) DNA replication through G-quadru-plex motifs is promoted by the Saccharomyces cerevisiae Pif1DNA helicase. Cell 145, 678–691

129. Johnson, J.E. et al. (2010) Altered gene expression in theWerner and Bloom syndromes is associated with sequenceshaving G-quadruplex forming potential. Nucleic Acids Res. 38,1114–1122

130. Mohaghegh, P. et al. (2001) The Bloom’s and Werner’s syn-drome proteins are DNA structure-specific helicases. NucleicAcids Res. 29, 2843–2849

131. Fay, M.M. et al. (2017) RNA G-quadruplexes in biology: princi-ples and molecular mechanisms. J. Mol. Biol. Published onlineMay 26, 2017. http://dx.doi.org/10.1016/j.jmb.2017.05.017

132. Sexton, A.N. and Collins, K. (2011) The 50 guanosine tracts ofhuman telomerase RNA are recognized by the G-quadruplexbinding domain of the RNA helicase DHX36 and function toincrease RNA accumulation. Mol. Cell Biol. 31, 736–743

133. Zhang, Y. et al. (2014) FMRP interacts with G-quadruplex struc-tures in the 30-UTR of its dendritic target Shank1 mRNA. RNABiol. 11, 1364–1374

134. Suhl, J.A. et al. (2014) Analysis of FMRP mRNA target datasetsreveals highly associated mRNAs mediated by G-quadruplexstructures formed via clustered WGGA sequences. Hum. Mol.Genet. 23, 5479–5491

135. Vasilyev, N. et al. (2015) Crystal structure reveals specific rec-ognition of a G-quadruplex RNA by a beta-turn in the RGG motifof FMRP. Proc. Natl. Acad. Sci. U. S. A. 112, E5391–E5400

136. Jain, A. et al. (2010) Human DHX9 helicase unwinds triple-helicalDNA structures. Biochemistry 49, 6992–6999

137. Booy, E.P. et al. (2016) RNA helicase associated with AU-richelement (RHAU/DHX36) interacts with the 30-tail of the long non-coding RNA BC200 (BCYRN1). J. Biol. Chem. 291, 5355–5372

138. Byrd, A.K. and Raney, K.D. (2015) A parallel quadruplex DNA isbound tightly but unfolded slowly by pif1 helicase. J. Biol. Chem.290, 6482–6494

139. Fekete, A. et al. (2012) The guanine-quadruplex structure in thehuman c-myc gene’s promoter is converted into B-DNA form bythe human poly(ADP-ribose)polymerase-1. PLoS One 7,e42690

140. Chakraborty, P. and Grosse, F. (2011) Human DHX9 helicasepreferentially unwinds RNA-containing displacement loops (R-loops) and G-quadruplexes. DNA Repair (Amst.) 10, 654–665

141. Thandapani, P. et al. (2015) Aven recognition of RNA G-quad-ruplexes regulates translation of the mixed lineage leukemiaprotooncogenes. Elife 4, 06234

142. Samatanga, B. et al. (2013) The high kinetic stability of aG-quadruplex limits hnRNP F qRRM3 binding to G-tract RNA.Nucleic Acids Res. 41, 2505–2516

143. McRae, E.K.S. et al. (2017) Human DDX21 binds and unwindsRNA guanine quadruplexes. Nucleic Acids Res. Publishedonline May 4, 2017. http://dx.doi.org/10.1093/nar/gkx380

144. Benhalevy, D. et al. (2017) The human CCHC-type zinc fingernucleic acid-binding protein binds G-rich elements in targetmRNA coding sequences and promotes translation. Cell Rep.18, 2979–2990

145. Norseen, J. et al. (2009) Role for G-quadruplex RNA binding byEpstein-Barr virus nuclear antigen 1 in DNA replication andmetaphase chromosome attachment. J. Virol. 83, 10336–10346

Trends in Biotechnology, October 2017, Vol. 35, No. 10 1013